THE MUSICAL INSCRIPTION FROM EPIDAUROS
(PLATE
14)
AN INSCRIPTION of the Imperial period which may add substantiallyto our understanding of ancient Greek music, SEG XXX, 390, was excavated northeast of the
palaistra at Epidauros in July, 1977. M. Mitsos published its text in 1980, printing the
presumedmusical notationwithout commentary.1In 1982 S. Sepheriadespresentedan interpretationof the notation based on Pythagoreanprinciples.2In a paper deliveredthe following year, J. Solomonofferedanotheranalysis of the musicalnotation,which he has now
published.3The inscriptionhas morerecentlybeen the subjectof an articleby M. L. West in
which he responds to Solomon's proposedinterpretationof the musical notation and discusses its contributionto our knowledgeof how hexameterswere sung.4In this articleI shall
firstpresentmy argumentsfor assigningSEG XXX, 390 to the same inscriptionas IG IV2 1,
135, followedby my text and commentary.I shall then proposeyet anotherinterpretationof
the musical notation,which I take to be a kind of musical shorthand.
THE FRAGMENTS
SEG XXX, 390 containseleven fragmentarylines of a hymn to Apollo and probablyalso to
the Muses.5 The top and right-handedges are extant. IG IV2 1, 135 was first publishedby
C. Blinkenbergin 1894.6 It was subsequentlypublished by M. Fraenkel as IG IV, 1185
1 M. Mitsos, ?e'IEpos iLvos 'e 'AO-KArJ'ITEL0v
'E7rtbabpov?, 'Apx'E4
SEG XXX, 390, edited by R. Stroud (1980).
1980, pp. 212-216
(=Mitsos);
Other works frequentlycited are abbreviatedas follows:
= Anonymus Bellerman, D. Najock, ed., Drei anonyme Traktate uber die Musik, Gottingen
1972
Blinkenberg = C. Blinkenberg,Asklepios og hans Fraender, I, Hieron ved Epidauros, Copenhagen 1894
H. Orph.
= Orphei hymni, W. Quandt, ed., 3rd ed., Berlin 1962, repr. 1973
= K. von Jan, ed., Musici scriptores graeci, Leipzig 1895, repr. Hildesheim 1962
Jan
AnonBell
Sepheriades = S. Sepheriades, <<'E
kpos '4lvos ke 'AO-K?X'LEL'Ov 'Ezrt8atpov>>,Eighth International Congress
on Greek and Latin Epigraphy, Athens 1982, 'AvaKoLvo-ELs', pp. 156-159
Solomon
= J. Solomon, "The New Musical Fragmentfrom Epidauros,"JHS 105, 1985, pp. 168-171
West
= M. L. West, "The Singing of Hexameters: Evidence from Epidauros," ZPE 63, 1986,
pp. 39-46
2 Sepheriades.I thank Lucy Meritt for making an abstractof this paper available to me. The second and
third volumes of the Proceedings have not yet appearedin print, and so I have been unable to consultthe full
text of Sepheriades'paper.
I Solomon. The abstractof his paper, which was read at the Annual Meeting of the ArchaeologicalInstitute of America, Cincinnati, 1983, is printed in AJA 88, 1984, p. 261.
4 West.
5 For the Muses, see line 11. Mitsos restoredreferencesto the Muses in both line 7 and line 8. To his
comparanda(Solon fr. 13 West and Krates fr. 1 Diehl [= H. Lloyd-Jones and P. Parsons, Supplementum
Hellenisticum, Berlin 1983, no. 359]) add the similar H. Orph. 76 Quandt and IG XII 7, 95 of Imperialdate,
from Amorgos.That the Muses were significantat Epidaurosin the Imperialperiod is attestedby a shrine to
them in the upper sanctuaryof Apollo Maleatas; see "Epyov1983, pp. 59-64.
6 Blinkenberg,no. 15, p. 126. Blinkenberg'sreading also appears in J. Baunack, "Zu den Inschriftenaus
Epidauros,"Philologus 54, 1895 (pp. 16-63), p. 19, no. 7.
American School of Classical Studies at Athens
is collaborating with JSTOR to digitize, preserve, and extend access to
Hesperia
®
www.jstor.org
52
SUZANNE BONEFAS
and, with revisions,by F. Hiller von Gaertringenas IG IV2 1, 135. It consistsof two small
stones (Fragments I and II), each of which is made up of six incomplete lines. In both
fragmentsthe bottom edge is extant, as is the left margin of Fragment I. This fragment
refersto severalmembersof Asklepios' "family":laso, Akeso, and Hygieia. FragmentII is
addressedto Apollo Maleatas and possiblyalso Asklepios,accordingto the restorationmade
by Fraenkel.
There are both epigraphicaland stylistic reasons for assigning SEG XXX, 390 to the
same inscriptionas the two fragmentsof IG IV2 1, 135. The findspotsare, unfortunately,
irrelevant,since all that is known about no. 135 is that the fragmentswere found in the
"hieron".The two inscriptionsare on the same type of darkred stone.7The threefragments
have all been sheared off so that the original back surface does not survive. It is therefore
uncertainwhether the stones were part of a stele or a block.That the latter is more likely is
suggestedby the similarityto IG IV2 1, 132-134, which consistof three hymns on the same
red stone and inscribed on what is certainly a building block. The preservedtop edge of
SEG XXX, 390 has anathyrosis,the stippling beginningca. 0.010-0.011 m. from the front
edge. The preservedbottom edges of no. 135 are worked in a similar manner but with a
broadersmoothband, ca. 0.030 m.
There is no reasonto doubtthat all three fragmentswere inscribedby the same mason,
who also inscribedIG IV2 1, 132-134.8 The letter forms and sizes are identical. Moreover,
the same diacriticalmark is used in both inscriptions,perhapsto indicatediaeresisor rough
breathing.9The fragmentsare also metricallycompatiblesince both SEG XXX, 390 and
no. 135, unlike the other Epidaurian hymns, were composedin dactylic hexameters.The
contentof the fragmentsargues furtherfor a connectionbetweenthem:SEG XXX, 390 is a
hymn to Apollo, IG 1V2 1, 135, Fragment I is addressedto Asklepios and Fragment II to
Apollo Maleatas.
Although there are not enough edges extant to be certain,the precedentof the Epidaurian hymns10suggeststhat SEG XXX, 390 and IG IV2 1, 135 were probablyalso arranged
in two columns. FragmentI would thus providethe left margin and bottomedge of the first
column, while SEG XXX, 390 preservesthe top edge of the secondcolumn as well as the
right edge of the inscription. Fragment II, in which neither of the side edges is extant,
completes the second column with the bottom edge. With this arrangement,the fact that
SEG XXX, 390 is in all probabilitythe beginning of a hymn indicatesthat there were at
least two hymns in this inscription:one to Asklepios in the first column and one to Apollo
(and Asklepios?) in the second. It is also possible that the second column containedeither
two differenthymns to Apollo, the first to Apollo Mousagetesand the secondto Apollo Maleatas, or a hymn to Apollo followed by one to Apollo and Asklepios. In favor of the latter
I
The editorscall it limestone;it is not, however,the same as the commonlocal pink limestone.
See below, commentaryon SEG XXX, 390, line 1, letter 10.
9 See below, commentaryon IG IV2 1, 135, Fragment II, line 4.
10IG IV2 1, 132-134, which have already been mentioned, contain hymns to Hygieia, Asklepios, and
Pallas. IG IV2 1, 129-131 contain hymns to All Gods, Pan, and the Mother of the Gods.
8
53
THE MUSICAL INSCRIPTION FROM EPIDAUROS
interpretation,the seemingly incongruouspresenceof the Muse at the end of the hymn to
Asklepios (M[do-]a,FragmentI, line 6) may be explainedas a transitionto the first hymn in
the secondcolumn, addressedto Apollo and the Muses.11 The first line of SEG XXX, 390,
which containsthe musical notation,might well have extendedacrossthe entire inscription
so that the music designatedby it applied to all the hymns on the stone.
SEG XXX, 390
H. 0.13 im., W. 0.15 im., Th. 0.04m.
IG IV2 1, 135, Fragment I
H. 0.07 m., W. 0.22 m., Th. 0.05m.
IG IV2 1, 135, Fragment II
H. 0.105 m., W. 0.21 m., Th. 0.05i.
SEG XXX, 390
]EvAZ 0 F= A v 'ao[tL
it]
aElo,ev
] avpron,otv
----------
1 [------]---------------------A-[--------?E]ov
[-------[------ 'A7r6?w]vt
{lacunaof ca. 18-25 lines)
5 [---------
]c
be
KXVTOT0o,6
a'
a&tow
KaTa?e4aL.
-------E7rtTTa]/LErvs9
Ka%L
Zvov 'O?vu7d]ov ayXaa TEKva
[MvacAooiVvas9
e
[-----------]Tov
10
Oot T[op]
[- - - - - - - - - - ] OVV'EKa7[
----
MO[-------
[---------]
{lacunaof ca. 5-12 lines)
IG IV2 1, 135, Fragment I
1
.. .. .
..Krrtp
.
--------Tp
[----[Ka]?JvKM ----[o-?]
6
ab'I
5
IG IV2 1, 135, FragmentII
-----]
. . . . . . . . . .. .__--.]Aara
\ [ ---------------?
?----------?
te ----[
'AKEOdTE[?]
[]oOat
[
avt8'a'vO' 'Yyelave'/ ?I---------]
a?a v EvoTVTot pe'va [ ?-----------?
XA[--------]
7J/,ETEp7Jt 8 %M[or]q,f'[7rel]
?------
II
1 [----------]
5
]
V----------]jEVs
?
tXa[ov?---------
?
?
]sEVEOVl/-----[
F
['A(rKX?A7Tru,Kat
'A7rA]AwOvKOVF0VEV?jU
[
]
]
[-]
]TpLs
I assume here that the three hymns were composedby a single poet who wrote them as an ensemble or
that a compiler chose them with thematic compatibilityin mind. Cf., e.g., IG IV2 1, 129-131, which share
themes and language.
54
SUZANNE BONEFAS
IG IV2 1, 135, FragmentI12
Line 1. IH Blinkenberg;IHI1Fraenkel;IHTHiller von Gaertringen.Fraenkel prints pi and restoresrl[atav]
even though his own drawing as well as the stone clearly identifiesthe letter as tau. Peek restores[, &'ETE]K
Line 2. Although Fraenkel bracketedthem, traces of both the initial lambda and the final mu are evident on
the stone. Blinkenberg did not read them. Hiller von Gaertringen dots the kappa even though it is clear
enough on the stone and in the drawings of the earlier editors.The restorationsare those of Fraenkel.
Line 3. The oblique strokeprecedingthe delta suggests alpha, not eta as read by all the editors.
Line 5. Of the final dottedalpha, the bottomof the left oblique is extant.
Line 6. After the first lacuna (a chip in the stone), only the right oblique of the dottedalpha is visible. I prefer
becausethe size of the lacuna suggeststwo ratherthan three missing letters. Both
M[do],a to Hiller's M[ot$o-]a
spellings are attestedat Epidauros.14
SEG XXX, 390
Line 1. Letter 3. A Mitsos, Sepheriades,Solomon;-AWest. West's emendationis made on purely musical
grounds.15
Letter 5. e Mitsos; e Solomon, followed by West. This emendationis also based on musical, not epigraphical,considerations."6
Letter 6. IGMitsos; F= Stroud; E Sepheriades, Solomon, West. West cites letter 1 of this line as a
but is in fact roundedlike the other
comparandum;however,that letter is by no means "noticeablysquare"'17
epsilons in the inscription,a fact which the brokenedge of the stone obscuresin Mitsos' photograph.I interpret this letter as a gamma with a diseme18followed by =,=whichI take as a punctuationmark.
Letter 7. A Mitsos, followed by Solomon and West; A Sepheriades.Neither the stone nor the squeeze
confirmsthe presence of the upper horizontal given by Sepheriades.A, an attested musical symbol, is more
likely.
Letter 10. A Mitsos, Stroud,followedby Solomonand West. What Mitsos interpretsas the bottomstroke
of a delta19comparesmuch more favorablywith this mason's alphas, which have a pronouncedupward tilt
from left to right. The bottomstrokeof his deltas is either almost horizontal(e.g., b, line 5; aotHjg,line 8) or
has a downwardtilt from left to right (e.g., be, line 9). IG IV2 1, 132-134, which were inscribedby the same
mason, show similar alphas and deltas.
Letter 11. E Mitsos, Stroud;0 West. Mitsos suggests the restorationEav8E[KaT
jL]. Solomon recognizes
that epsilon is "extremelyquestionable",20but he does not suggest an alternative.West reads the omikronas
12 My original readings (presentedin my M.A. thesis, University of Texas, 1986) were based on photographs and squeezes kindly furnishedby Paula J. Perlman,who also drew my attentionto the possibilitythat
these fragmentsmight belong to the same inscription.I have since examinedthe stonesmyself and made a few
changesto my original text.
13 W. Peek, Inschriftenaus dem Asklepieion,AbhLeip 60, fasc. 2, 1969, no. 59, p. 55. Cf. H. Orph. 67
Quandt, line 1.
14 Moi3o-a:
IG IV2 1, 128, line 40; 541, 542, 543. Moo-a(-M o-a): Peek, AbhLeip 63, fasc. 5, 1972, no. 1
(Moo-ayE'Ta);SEG XXXIII, 304. Moitoais also attested:IG IV2 1, 129, line 7.
15 West, p. 41. He makes this emendation in order to avoid changing tetrachordswithin the chromatic
Hyperionian tonos. For a more detailed discussionof the musical symbols,see pp. 56-58 below.
16 The emendationallows Solomonto place all the notes, with the exception of the next, within one tonos,
the chromaticHyperionian.
17West, p. 41.
18 A diseme is a musical symbol of durationwhich indicatesthat a syllable is to be sung for two counts. See
footnote31 below.
19
Mitsos, pp. 213-214.
20
Solomon,p. 170.
THE MUSICAL INSCRIPTION FROM EPIDAUROS
55
dottedin the belief that sigma and epsilon are also possible.The upward returnof the lower curveof the letter,
however,extends too far for a sigma or an epsilon, and there is no traceof a centralhorizontal,thus ruling out
theta.
Lines 2-9. Unless otherwise noted, the restorationsare those of Mitsos.
Line 7. I see no trace on the stone of the initial dotted iota read by Mitsos and Stroud and printed as
certain by Solomon and West. It remains, however, the most likely restoration.Solomon, followed by West,
has restored 'Ao-KAX7'z[L0ov].
Line 8. As Stroudhas pointedout, the initial sigma is mistakenlybracketedin the editionof Mitsos. Two
dots appearover the upsilon in V'`lvos.This diacriticalmarkoccursover severaliotas in the Epidaurianhymns
but above upsilon only here.21These dots are not read by Mitsos.
Line 9. Only the right tip of the horizontalof the initial tau is legible. There is room at the end of the line
for only two relativelynarrow letters.
Line 11. MO Mitsos, Stroud;ME West. I do not know the groundsupon which West bases his reading. The
stone clearly reads omikron.M[uor]a ( or some form thereof) should perhaps be restored.
IG IV21, 135, Fragment II
Line 2. I dot the initial alpha because, epigraphically,the presenceof the intersectionof the oblique and the
horizontaldoes not rule out the deltas of this mason.22Of the final tau the left tip of the horizontalis extant.
Line 3. I can see no trace on the stone of the initial tau, printed as certainby Hiller.
Line 4. The theta, not read by Blinkenberg,is faint but certain.The precedingdottediota mightjust as easily
be upsilon. To the left of this strokeI read a curvedstrokeof which the lower tip of the returnis extant. I print
sigma, although epsilon can not be ruled out. The iota of zfpo- has two dots above it, as Hiller notes. This
notationcommonlyoccursover iota, less frequentlyover upsilon, in a numberof other Epidaurianinscriptions
of the Imperial period. It is usually understoodas an indicationof diaeresis,but perhaps in some instancesit
denotedspiritus asper, e.g., in vlAvos'(SEG XXX, 390, line 8) where diaresis is not possible.23The O(Eatare
perhaps the Muses.24
Line 5. Hiller reads the initial vertical stoke as a dottediota. Two dots over the iota of zAa[- -- ] are noted by
this editor but not by his predecessors.Following the final alpha, Hiller reads PO (undotted).There are no
traces of these two letters on the stone, and neither earlier editor noted them. Perhaps they were mistakenly
read from the end of the previous line, an error compoundedby the editor'ssuggestedrestorationof 'I?aipos
here. That this inscriptioncloses with a prose postscripthas been noted by all the editors.25The hexameters
definitelyend by line 6. Hiller, however,reads this line as the beginningof a dedication,taking 'IAa[pov] and
in the next line as names.26
E4vIIE'vovs
a restoration which fits the traces well.27
Line 6. Peek reads Tt4- - -] following EV/LEVOVg,
Line 7. Enough of the initial lambda is present to follow the editorsof IG IV in printing it as certain. Of the
final sigma, the lower curve is extant. F occurs in a similar context in IG IV2 1, 133, lines 14 and 15. As the
editorsof IG IV have pointed out, it indicatesa triple repetitionof some sort. Precisely how much of the text
was to be repeated,however, is unclear.
21 See below, commentaryon IG
IV2 1, 135, Fragment II, line 4.
For this mason's alphas and deltas, see above,commentaryon SEG XXX, 390, line 1, letter 10.
23 In Epidaurianinscriptionsof the Imperial period,spiritus asper is indicatedby 1--and'
(IG IV2 1, 418
and 419) and perhaps also by (IG IV2 1, 420).
24
See footnote5 above.
25 The editors of both editions of IG IV and K. Munscher, "Zu den neuen Hymnen aus Epidauros,"
PhilWoch 52, no. 35-38, 1932, cols. 87-102.
26 As Hiller and Munscher (op. cit., col. 91) have pointed out, the name Hilaros is known from two dated
inscriptionsfrom the Maleatas sanctuary,IG IV2 1, 393 and 400.
27
Peek, footnote 13 above, no. 59, p. 55.
22
SUZANNE BONEFAS
56
Line 8. TOIL Fraenkel, Hiller von Gaertringen;TPIY Blinkenberg, Peek.28The omikrons of this mason
elsewhere always occupy the whole letter space, not just the upper portion of it. Rho is the only possible
reading.
THE MUSIC
The commentatorsdisagreeabout which of the symbolsin the first line of SEG XXX, 390
should be interpretedas notes. Accordingto Sepheriadesand West, all eleven charactersin
the firstline shouldbe interpretedas musicalsymbols,which are correlatedwith the wordsin
the line directlybelow them, in accordancewith the typicalconventionsof ancientGreekmusical notation.29The final four letters,West believes,are instrumentalnotation.Solomon,on
the other hand, although he entertainsthe possibility that the last four letters are notes, is
moreinclinedto believethat only the first sevenlettersare musicalsymbolsand that the final
four are marginal notation.30I shall argue that only the five letters following the initial EN
are musical symbolsand that they were deliberatelyset off fromthe rest of the text.
Each of the five letters in question has above it the rhythmicsymbol called a diseme.31
The lack of such marks over the other letters does not in itself indicate that they are not
notes, since a single rhythmicunit is regularlydenotedby a lack of rhythmicnotation. It is,
however,sometimesthe case that the diseme is used relatively,so that "theunit of measurement is not a short, but a long syllable."32In moderntranscriptionsof Greek musical documents, rhythmicnotationis often so treatedfor the sake of convenience.
Not only are the five letters in question more widely spacedthan the other letters in the
inscription,but there is also a larger than averagespace separatingthem from those which
precedeand follow them: 0.008 m. between the nu and the alpha and 0.012 m. between the
"lambda"and the epsilon. The widest spacing within the body of the inscriptionis between
the initial alpha and the tau in KaraXA'eat
(line 6), which are separatedby 0.007 m. In contrast, the distancebetween the alpha and the zeta in line 1 is 0.009 m., a considerablywider
spacing. The spacing between the rest of the five symbolsis comparable:0.008 m. between
the zeta and the theta, 0.007 m. between the theta and the gamma, 0.009 m. between the
gamma and the "lambda".The space between the initial epsilon and the nu, however,is the
same as that between the final epsilon and the nu in line 2 and is also equal to that between
28
Ibid.
For a brief introductionto ancient Greek music and musical notation, see the entry by J. F. Mountford
and R. P. Winnington-Ingramin OCD2,s.v. Music and the entry by Winnington-Ingramin the New Grove
29
Dictionary of Music and Musicians, Washington, D.C. 1980, s.v. Greek Music, Ancient.
30 Solomon, p. 170. Although he considersMitsos' bTVE[KaTijt] probable,he also suggests that these letters
may be part of a "melismaover the -MAeV
syllable"or instrumentalnotation.He recognizesthe difficultywhich
"theliberal spacing between the lambda and the ensuing epsilon"presentsfor the formersuggestion.
31 AnonBell III.83, p. 138 in the edition of Najock. The four rhythmic signs given by AnonBell are
and 'L', representingtwo, three, four, and five counts,respectively.Only the first two are known
'-(-),"-,
from extant musical documents,and indeedthe secondappearsin a formwhich is a mirrorimage of the siglum
in AnonBell.
32J. F. Mountford, "GreekMusic in Papyri and Inscriptions,"in New Chaptersin the History of Greek
Literature,ser. 2, J. U. Powell and E. A. Barber,edd., Oxford 1929 (pp. 146-183), pp. 157-158.
THE MUSICAL INSCRIPTION FROM EPIDAUROS
57
the epsilon and the nu which follow the musical notation in line 1, further support for the
identificationof the eighth and ninth charactersas lettersratherthan notes.33
The musical symbols also tend to be larger than their counterpartsin the body of the
inscription,especiallythe alpha, theta, and "lambda".The theta in line 1 is 0.004 m. wide,
the alpha 0.005 m., and the "lambda"0.004 m. The theta in line 3 is 0.003 m., the alpha in
line 5, 0.004 m., and the lambdasin lines 4, 6, and 7 are 0.003 m., 0.003 m., and 0.004 m.
wide respectively. The average difference between the musical symbols and their nonmusical counterpartsin both SEG XXX, 390 and IG IV2 1, 135 is ca. 0.001 m.
The interpretationof three of the musical symbols, 0, F=, and A, is problematic.As I
noted in the epigraphicalcommentary,Solomon is followed by West in reading 0 for the
fifth letter, despite the fact that the crossbardoes not extend beyond the perimeterof the
circle. Solomon is convinced, in part because of his reading of the sixth letter, that the
musical symboltheta would require an unprecedented"doublemodulation".By attributing
the inscribing of theta for recumbentphi to a "musicallyuntrained stone carver,"he can
assign the notes as he reads them, except for the seventh symbol, to the chromaticHyperionian tonos.34Nevertheless, theta is a legitimate musical symbol which does, as I shall
demonstrate,make sense musically in this context.
The sixth letter in line 1 is the most problematic.Sepheriades,Solomon,and West read
epsilon. All three interpretthis letter as representingthe same note as the first letter in the
line. Sepheriadesand West also interpret the eighth letter in this way, and Solomon too
suggests the possibility that it is further musical notation. It is, however, epigraphically
doubtfulthat either the first letter or the eighth representsa note. That these, like the other
tepsilonsin the inscription,are lunate ratherthan straightis evidenton the stone.
There are further difficultieswith reading letter 6 as an epsilon. The crossbarsof the
other epsilons in this inscriptionare never detachedas they are in this letter and neverslope
upward to the right to such a degree.West implies that the shape of this letter was intended
by the writer of the exemplar to distinguish it from other, non-musical, epsilons in the
inscription,but that the stonemason,ignorantof musical notation,was unsure of the identity of the letter. Assuming that the exemplar was consistent, this argument nevertheless
requiresthat the first and eighth letters in line 1 be regardedas somethingother than notes,
an interpretationwhich is compatible neither with West's reading nor those of the other
commentators.Moreover, if the mason had set out to distinguish this "epsilon"from the
others, why did he not go to similar lengths with the other notes?
Stroud plausibly reads this letter as F=, and a comparisonwith the gamma of ayXaa
(line 7) leaves no doubt about the identificationof letter 6 (minus its "crossbars")as a
gamma, followed by a siglum consistingof two parallel horizontalbars. Stroud, however,
33 The average letter width for lines 2-11 is 0.005 m., as opposed to 0.008 m. for the musical symbols and
0.007 m. for the first line in general. The averagespace between letters in lines 2-11 is 0.003 m., contrasted
with 0.006 m. for the musical symbols. The first two and the final four letters of line 1, however, average
0.005 m. in width and 0.003 m. in inter-letterspacing.
34 Solomon,pp. 169-170. The note thus standsas Nete in the tetrachordSynemmenonof the chromaticHyperioniantonos.
58
SUZANNE BONEFAS
does not note the diseme to the upper left of the letter, the placementof which has parallels
in other musical texts.35Although there is no parallel for its use in a musical context, the
siglum may have servedas some sort of punctuationmark. Strouddoes not offer an interpretationof the symbol, which is in any case relativelyrare. In most of the inscriptionsin
which it appears,its functionis unclear.36A full stop, however,would be appropriatein this
position since, as I shall argue below, gamma is the last note beforea rest.
The next symbol, A, also presents difficulties.It is interpretedby Solomonas the note
lambdawith a diseme. Because lambdadoes not occurin the chromaticHyperioniantonos,
Solomon offers three interpretations:it must be understoodas a "borrowednote", or as
"anothererror on the part of the stone carver"whereby lambdawas inscribedfor alpha or
delta, or the symbol is a pointed leimma.37Sepheriadesand West both interpretit as Solomon's third alternative. Sepheriadesreads A, which he calls a three-countrest, although
accordingto Anonymus Bellerman (our only theoretical source for rests38)a rest of this
durationwould be depictedas A (or A).39Furthermore,there is no evidenceon the stone for
the upper horizontalwhich Sepheriadesrecords.I thereforefollow West in interpretingthe
symbol as A, Anonymus Bellermann'ssign for a rest of two-count duration, attestedin a
numberof musical documents.40
I turn now to the rest of the line and the relation of the non-musicalletters (the initial
EN and the final ENAO) to the musical notation. In all the extant musical texts, as I have
mentioned,musical symbolsare placed directlyover the syllable on which the note in question is to be sung; thus the notationoccursthroughoutthe text. The positionof the notation
in SEG XXX, 390 is therefore anomalous:it occurs only in the first line and there not
centered over any particular syllables. The view of Sepheriadesand Solomon that these
notes should be associatedwith the syllables below them is improbablebecause there is no
evidencefor hymns of which only one line (or part of a line) was sung and the rest declaimed
without accompaniment.41West, like Sepheriadesand Solomon, begins by assuming that
35 Cf. Mesomedes, Hymns 1 and 2 (E. Pohlmann, Denkmdler altgriechischeMusik, Nuremberg 1970,
nos. 1 and 2), in which the copyist appearsto have positionedthe disemeto one side over notes which, like this
one, have upper horizontals.
36 R. Kaiser in De inscriptionumgraecarum interpunctione,Berlin 1887, does not mention this siglum.
Parallel horizontal lines may also represent single strokes in numerical notation (cf., e.g., IG IV, 481 from
Nemea). If such is the case here then perhapsthe siglum indicatesthat the four notes are to be repeatedtwice.
The resultingtwelve-notephrasewould be ideal for a hexameterhymn. Such an interpretationwould perhaps
support West's argumentsfor a single phrase repeatedfor each line of the hymn. See p. 59 below.
37 Solomon, p. 170. Although he mentions the possibility that this letter is a leimma or rest, he prefers to
considerit a note.
38 AnonBell 111.102, p. 150 Najock.
39 The latter sign is given by Anonymus Bellerman;however, his three-countdurationalsiglum (triseme),
I,always occursin extant texts as '. There is no example of a three-countrest in the extant fragments,but
such a reversalbetween theory and practicemight hold for rests as well as durationalsigns.
40
In its rounded and pointed forms it occurs in several instrumental pieces from AnonBell (P6hlmann
[footnote35 above],no. 10) and the vocal POxy 1786 (Pohlmann,no. 34), POslo 1413 (Pohlmann,nos. 36 and
37), and PMich 2958 (Pohlmann,nos. 39 and 40).
41 Solomon assumes from the outset that the musical notation is correlatedonly with the text immediately
below it, although he recognizes "the careless positioning of the alleged musical letters" (p. 169) and the
"puzzling"fact that &EaojMev "mightbe the only musically notatedword in the hymn"(p. 171).
THE MUSICAL INSCRIPTION FROM EPIDAUROS
59
the first seven letters of line 1 are correlatedwith the syllables below them in the usual
manner. He goes on, however,to speculatethat this one-line melodywas repeatedfor each
line of the text. Thus West sees in this inscriptionan actual example of the practicehe had
previously postulated for the Homeric aotbot',
namely, a repetitiousmelody punctuatedat
the end of each verse by a short instrumentalinterlude.42In the presentpiece this interlude,
in West's interpretation,is representedby the final four lettersof the line.
If, as I have argued, the notes did not extend across the whole line in such a way that
they could be repeated for each subsequentline, then there remain two relationshipsto be
explained:that between the four notes (plus a rest) and the rest of the line and that between
the first line and the rest of the hymn. It is possible that the musical symbolswere intended
as some sort of "shorthand",presumablyfamiliar to the performer,from which he could
determineeither the tonos which was designatedfor the melody or the standardliturgical
tune to which he was to set the hymn. The line as a whole may have served,in effect, as a set
of instructionsto the performerand must have given him enough informationto performthe
entire hymn. At the end of the line, I restoreczvao[tbijt]which fits both the context and the
available letter spaces.43Thus the text which precedes and follows the musical notation
must have clarifiedthe purposeof the notes, signifyingsomethingto the effectof "singto the
tune (or tonos) of X in the song."
If what we have in the first line of SEG XXX, 390 is indeed an example of Greek
musical shorthand,then this piece is without parallel among extant musical texts.44The
fact that such a shorthandwould have been of interest primarily to performersmight account for the lack of referenceto it in the ancientmusicaltreatises,whose authorswere more
concernedwith theory than with applicationor performance.45
In his article on Homeric singing, West discussesthe musical significanceof the term
avapoxr, which may be relevant to the music of the present inscr'iption.46
In Homer, the
term refers to the initial, "tuning-up"strains of the lyre, and from Pindar we know that in
choralperformances"thepreliminarynotes of the lyre serveas a signal and guide to dancers
and singers."47West continues:
Elaborateava,oXat' seem to have been a featureof the dithyrambsof poets such as
Cinesias. Later writers use avafakXXEa-Oat
rather loosely of beginning a song, etc.,
with or without accompaniment.The Homeric ava,8oX does seem to have been
something more, involving some prefatory notes or strumming on the phorminx
beforethe singing began, but it was probablynothing extended.48
42
M. L. West, "The Singing of Homer and the Modes of Early Greek Music,"JHS 101, 1981, pp. 113-
129.
Cf. Plato, Rep. 399C:ev Tatg 4ba TE Katte` \eo-tv; Laws 664a: "' Tf waZ Kat Mt6otSL
Ki A
There are certainly parallels from other cultures for musical abbreviation,especially for sacred music.
We might compare,for example, the early (7th centuryafter Christ) notationof plainsongby meansof neums,
a system which was of no use to a singer not already familiar with the music.
45 A possible exception is Aristoxenus of Tarentum, who professed, at least, to have treated music as it
existed in practice.
46 West (footnote42 above), pp. 122-123.
47 Ibid., p. 122: Homer, Od. I.555 (VIII.266) and XVII.26I-263;
Pindar, Pyth. I.I-4. See West's notes 56-58
for a completelist of literary sources.
48Ibid.,p. 122.
3
44
60
SUZANNE BONEFAS
West goes on to speculate that such "instrumentalflourishes"occurredat the end of each
hexameter,but for the purposesof this study, it is the conceptof an initial, short instrumental passagewhich acted as a signal to performersthat is important.That an ava8oxArneed
not have been long or elaboratewe know, for example, from a scholiaston Pindar, 01. ix
Archilochos,the story goes, having
who relates the origin of the onomatopoeicrvEvAXXa.49
no kitharist to perform his hymn to Herakles, "imitatedthe tune with speech"(bta rtvos'
as a beginning for
XAeos' ro ufXosgCLM?j-aro)."Thus he drew up the phrase rvEvAXXa
in the midstof the
what followed,and he himself spokethe tune of the kithara,the rvEvAXXa,
chorus,and they did the rest."50Archilochoseither sang the kitharapart without words, or,
more likely, the word "tinella"("ting-a-ling")to the kithara part. Although I cannot conclude that the four notes.in SEG XXX, 390, line 1 constitutean ava/3oX j, it is relevantthat
there existed a traditionof signaling to performersand that perhapssuch preliminarynotes
informedthe performer(s)of the tune or tonos to be employed.
West raises anotherpoint about the Archaicpracticeof singing hexameterswhich may
throw some light upon the form of the "shorthand"of SEG XXX, 390. The lyre may
originally have had but four stringswhich presumablycorrespondedto the four notes of the
tetrachord.51Thus, West argues, Homeric hexameters were sung on four notes only, in
accordancewith pitch accent. It may thereforebe significantthat the abbreviationof this
hexameter tune consists of four notes. This is not to say that there is a tetrachordwhich
comprises these four notes. The number of notes involved, nevertheless,may be another
reflection of concern for tradition on the part of whoever was responsible for having the
hymn inscribed.However long a Homeric avap8oX might have been, it would have consisted of only four differentnotes, and the kitharodeneed only know those four notes to tune
his instrumentaccordingly.Even in later times, after the compassof the lyre had increased,
it might have remained traditional in hexameter poetry to use a four-note introductionto
acquaint the performerwith the tune. In any case, since the present piece undoubtedly
containedmore than the four notes inscribedin line 1, which have a total range of only 11/2
scale tones, and since the lyre in later times had more than four strings, the number four
does not seem to have had any practicalsignificance.In summary,although I do not agree
with West's reconstructionof the piece, I accepthis argumentthat it was the work of someone concernedwith preserving(or revertingto) traditionalmusical practice.
Since SEG XXX, 390 contains only four possibly non-sequential notes, which may
constitutenot a musicalpiece but an abbreviationof one, a definitiveanalysisof the music is
unlikely to be achieved.I shall begin with a brief outline of the basic principles of ancient
Greek music followed by a summary of Solomon'sand West's analyses, even though my
readingof the notes does not allow me to accepttheir conclusions.
49 Schol. 01. ix,
line 1 = Drachmann I, p. 266.
OVTW Ta 4D aVEfa?AAETO,
OVV TOVTO TO KO/4Ia TIVE?Xa,
EVMeOQ)TO Xop( e\eye, TO T7)vvekha, ? be XoP'o Ta d;\ot7ra.
50 OvvJTcag
Kat avTog WeV TO /.AAo
T^
KOapav
51 This assumption is based on literary evidence (e.g., Strabo, XIII.2.4;
and on
Cleonides [Jan], 202.II-I2)
archaeologicalevidencefrom the early Archaicperiod. On the other hand, seven-stringedlyres were depicted
in Minoan and Mycenaean art, and an anonymousrefereehas arguedthat this instrumentwas in continuous
use fromthe BronzeAge and that Geometricdepictionsof four-stringedlyres reflectstyle ratherthan reality;I
do not know how this view would accountfor the literaryevidence.
THE MUSICAL INSCRIPTION FROM EPIDAUROS
61
The most basic unit of ancient Greek music was the tetrachord,of which the two extreme notes are fixed while the two interiornotes vary accordingto genus (diatonic,enharmonic, chromatic).These tetrachordswere arrangedby theorists in such a way as to form
continuoustonoi ("pitch-keys")which were designedby variousethnic labels (e.g., Dorian,
Hypolydian, Hyperphrygian). Such arrangementsof tetrachordswere divided into two
systemata.The Greater Perfect System consists of the tetrachordsHypaton, Meson, Diezeugmenon, and Hyperbolaion (lowest to highest), the Lesser Perfect System of the tetrachords Hypaton, Meson, and Synemmenon.The Immutable System combines these two
systemata.The notes within the tetrachordswere designatedby position names which reflect the kithara strings on which they were played. Thus the notes of the two lowest
tetrachords,Hypaton and Meson, were called, in ascendingorder, Hypate (highest string,
not pitch), Parhypate,Lichanos,and Mese. Scale steps within the tonoi were designatedby
the position name followed by the name of the tetrachord(e.g., ParaneteSynemmenon).
As noted above (p. 54), Solomon,for musical reasons, reads letter 5 not as theta but as
recumbentphi. He finds it problematict-hatnowhere in Alypius' tables52does theta occurin
the same tonos as all the other notes in the line. Recumbentphi, however, does occur with
the other notes, as Solomonreads them, with the exceptionof letter 7 (if it is read, as it is by
Solomon, as a note rather than a rest), in the chromaticHyperionian tonos. West for the
most part agrees with Solomon's analysis, particularlywith his emendationof the theta,
except that he would prefer to remain as much as possible within a single tetrachord
(Synemmenon)and not to ventureinto the tetrachordDiezeugmenon,as Solomon'sreading
does. He therefore emends the third letter, alpha, to delta in spite of its epigraphicalcertainty. Both commentatorsnaturally assumethat the notes as they read them were intended
to be sung in direct sequence, and they must thereforemake emendationsin order to make
sense of the text. Given the sense of the text as I read it, however,it is no longer necessaryto
think of these notes as the tune itself but rather as representativeof it in some manner.
No single tonos containsall four notes of SEG XXX, 390 in their unemendedform, so
modulationis apparently involved. Cleonides (2nd century after Christ) in his treatise on
music distinguishedfour types of modulation: 1) KaTa ye'vos; 2) KaTa o-v'o-T?)/a;3) KaTa
The modulationof the presentpiecemustfirstbe considrovov; and 4) KaTa /.LEXo7ro0tav.53
ered as an example of the third type, since the four notes are not containedtogetherin any
one tonos, regardlessof genus or systema. Concerningthis type of modulation,Cleonides
further distinguished between the two types, "melodious"and "unmelodious"('K,ueXO's.),
"accordingas the notes which coincidein pitch are similaror dissimilaras regardstheir participation in the Pyknon."54In a tetrachorda pyknon occurs when the sum of the two
smallest intervals is less than the remaining interval;the two small intervals are then referred to collectivelyas a pyknon. Such a distributionof intervals may occur in the enharmonic and chromaticgenera.
4 [Jan], no. VIII,
52 Alypius, E'Lraywy%MOVG-LK
pp. 362-406.
%
4 [Jan], pp. 204-205. Other musical treatises give similai accounts of
Cleonides, ELcraywy 'APUOVLK
modulation.
54Jan, p. 205, translatedby H. S. Macran, The Harmonicsof Aristoxenus,Oxford 1902, p. 262.
SUZANNE BONEFAS
62
The following discussionis based on H. S. Macran's analysisof Cleonides'definitionof
modulation as outlined above.55According to Macran, the more two tonoi coincide, the
more harmonious is any modulation between them. The "most melodious"modulations
involve tonoi separated by a perfect fourth or a perfect fifth. "Less melodious"but not
unmelodiousare modulationsinvolvingtonoi separatedby one tone or five tones. All others
are "unmelodious".The conclusionderivedfrom Macran's analysis is that two tonoi separated by a perfect fourth or a perfect fifth share many common pitches, and modulation
between them is thereforevery melodious.Tonoi separatedby one or five tones share one
commonnote which holds the same position in the pyknon. Modulationbetween these keys
is thereforealso melodious,although to a lesser degree.
An examinationof the possibilitiesrevealsthat, given the notes of this inscription,there
can be no "mostmelodious"modulation.Indeed,only one possibilityis at all melodious.The
Hyperionian tonos contains alpha as Parameseand zeta as Mese and Nete Hyperbolaion.
One tone lower is the Lydian tonos, which has theta and gamma in the tetrachordSynemmenon, as well as the commonzeta as Paramese.The intervalAZ, a downwardprogression
from Parameseto Mese in the Hyperioniantonos, is also, in effect, the point of modulation
into the Lydian tonos, in which zeta occupiesthe same positionas alpha in the Hyperionian,
i.e., Paramese.Two types of modulationappear to be involved:the modulationKara ToVoV
from the Hyperionian to the Lydian tonos, and the modulationKara o-vo-T?)/.a, since alpha
and zeta occur in the Greater Perfect System of the Hyperionian tonos while theta and
gamma occur in the Lesser Perfect Systemof the Lydian tonos.
However the music of this fragment is interpreted,its importancenot only for Greek
music but also for Greek religion should not be underestimated.The linking of SEG XXX,
390 with IG 1V2 1, 135 yields a third set of inscribedhymns from the sanctuary at Epidauros.The presenceof musical notationprovidesconcreteevidencefor the performanceof
liturgicalhymns and so gives a glimpse of the workingsof cult ritual in the Imperialperiod.
The existence of a means of musical abbreviationsuggests that sacredmusic was probably
standardizedin this period. The apparentlyarchaic featuresof this notation perhaps indicate that such a processof standardizationhad alreadybegun beforethe Imperial period. It
may thereforebe possible to gain some insight into earlier liturgical music and poetry by a
judicious comparisonwith the later material. Finally, it may also be possible to apply to
other musical texts not only Cleonides'principleof modulationbut perhaps also, if caution
is exercised,other principlesderivedfrom theoreticaltreatises.56
SUZANNE BONEFAS
UNIVERSITY
OF TEXAS AT AUSTIN
Departmentof Classics
Austin, TX 78712
55For the entire argument,see Macran, op. cit., pp. 262-266.
56 On the pitfalls and rewards of such analyses and for an exemplary prototype, see J. Solomon, "The
Seikilos Inscription:A Theoretical Analysis,"AJP 107, 1986, pp. 455-479.
4-.~~~~~~~~~~~~~~~~~~~~
44
i4.
IC IV' 1, 135, Fragment I
IC TV' 1, 135, Fragment II
SUZANNE BONEFAS: THE MUSICAL INSCRIPTION FROM EPIDAUROS